| CPC G06T 7/55 (2017.01) [G06F 18/2148 (2023.01); G06T 7/70 (2017.01)] | 10 Claims |
|
1. A method for training a neural network for controlling a robot, comprising:
ascertaining a camera pose in a robot cell and an uncertainty area around the ascertained camera pose;
ascertaining an object area in the robot cell, which includes positions for an object to be handled by the robot;
generating training-camera-images, including, for each training-camera-image of the training-camera-images:
establishing a training-camera-image camera pose randomly in the uncertainty area around the ascertained camera pose;
establishing a training-camera-image object pose randomly in the object area;
generating the training-camera-image, so that it shows the object with the training-camera-image object pose from the perspective of a camera with the training-camera-image camera pose;
generating training data from the training-camera-images, each of the training-camera-images being assigned one or multiple training-robot control parameters for handling the object in the respective training-camera-image object pose of the training-camera-image; and
training the neural network using the training data, the neural network being trained to output, from a camera image that shows the object, a specification of one or of multiple robot control parameters for handling the object.
|